A0189
Title: Robust statistical inference for one-shot devices
Authors: Elena Castilla - Universidad Rey Juan Carlos (Spain) [presenting]
Abstract: One-shot device testing is an increasingly important problem in the area of reliability. This is an extreme case of interval censoring, where one only knows if the device works when it is tested. The existing literature on one-shot devices is extensive and focuses particularly on the development of techniques for estimating the maximum likelihood estimator (MLE). The MLE however, presents an important lack of robustness in the presence of outliers. We present alternative divergence-based estimators, which are seen to be more robust with an unavoidable loss of efficiency.